linkify: handle URLs in parenthesized statements

Sometimes, URLs exist at the end of parethesized statements,
and we shouldn't unnecessarily capture that.

(example: https://public-inbox.org/ruby-core/20170623032722.GA8124@dcvr/)

diff --git a/lib/PublicInbox/Linkify.pm b/lib/PublicInbox/Linkify.pm
index 8e1728c71136c78de06af4d53603e54116496edb..93c468fe74839c6d2a0e818996e8a00680a70464 100644 (file)
--- a/lib/PublicInbox/Linkify.pm
+++ b/lib/PublicInbox/Linkify.pm
@@ -25,7 +25,7 @@ my $LINK_RE = qr{(\()?\b((?:ftps?|https?|nntps?|gopher)://
 sub new { bless {}, $_[0] }
 
 sub linkify_1 {
-       $_[1] =~ s!$LINK_RE!
+       $_[1] =~ s^$LINK_RE^
                my $beg = $1 || '';
                my $url = $2;
                my $end = '';
@@ -41,6 +41,8 @@ sub linkify_1 {
                        }
                } elsif ($url =~ s/([\.,;])\z//) {
                        $end = $1;
+               } elsif ($url !~ /\(/ && $url =~ s/\)\z//) {
+                       $end = ')';
                }
 
                # salt this, as this could be exploited to show
@@ -51,7 +53,7 @@ sub linkify_1 {
                $url =~ s/&/&/g;
                $_[0]->{$key} = $url;
                $beg . 'PI-LINK-'. $key . $end;
-       !ge;
+       ^ge;
        $_[1];
 }
 

diff --git a/t/linkify.t b/t/linkify.t
index 99acf17d593830884b1533452bf2ceae9ef8cd28..041c15c73144efcd8c966ce08f71f7ed10c9e94f 100644 (file)
--- a/t/linkify.t
+++ b/t/linkify.t
@@ -14,6 +14,16 @@ use PublicInbox::Linkify;
        is($s, qq(<a\nhref="$u">$u</a>.), 'trailing period not in URL');
 }
 
+# handle URLs in parenthesized statements
+{
+       my $l = PublicInbox::Linkify->new;
+       my $u = 'http://example.com/';
+       my $s = "(see: $u)";
+       $s = $l->linkify_1($s);
+       $s = $l->linkify_2($s);
+       is($s, qq{(see: <a\nhref="$u">$u</a>)}, 'trailing ) not in URL');
+}
+
 {
        my $l = PublicInbox::Linkify->new;
        my $u = 'http://example.com/url-with-trailing-semicolon';